We propose (1) to develop and test a model of visual interest-maintenance (or perceptual momentum) in children and adults, using motion pictures and printed text; (2) to extend and refine an overall account of cinematic representation and visual comprehension within the framework of a general multi-level theory of perceptual organization, and to provide measurement techniques and data at important points in that account. The work in cinematic representation is necessary for the study of momentum, but it is also important in its own right. Experiments are described in varying detail, in the context of a general program: Subjects view short films (and read text) about objects, scenes and events, while direct and indirectmeasures of comprehension and interest-level or momentum are obtained. The data will test and refine the models of comprehension and of momentum. Replications and cross-validations will be performed where appropriate.